This document has been converted from the original publication:
`Barry, K. M., Cavers, D. A. and Kneale, C. W., 1975, Report on recommended standards for
`digital tape formats: Geophysics, 40, no. 02, 344-352.
`
`
`
`RECOMMENDED STANDARDS FOR DIGITAL TAPE FORMATS1
`
`K. M. BARRY2, D. A. CAVERS3, AND C. W. KNEALE4
`
`It is assumed that this format will accommodate the
`majority of field and office procedures and the
`techniques presently utilized.
`FORMAT SPECIFICATION
`The following general information describes the
`recommended demultiplexed format (Figure 1):
`1) Tape specifications,
`track dimensions and
`numbering, and all other applicable specifica-
`tions shall be in accordance with IBM Form GA
`22-6862 entitled "IBM 2400-Series Magnetic
`Tape Units Original Equipment Manufacturers'
`Information".
`At the present time, IBM has proposed an American
`National Standard for the 6250 CPI group coded
`recording format. Should this format be used within
`the geophysical
`industry,
`the applicable
`IBM
`specifications would apply. The additional formatting
`required by this proposed method is a function of the
`hardware and thus becomes transparent to the user.
`2) Either the NRZI encoded data at 800-bpi density,
`or the phase encoded (PE) data at 1600-bpi
`density may be used for recording.
`3) All data values are written in two’s complement
`except the 320bit floating point format, Figure 3-
`A, which is sign, characteristic, and fractional
`part.
`4) Data values are written in eight-bit bytes with
`vertical parity odd.
`
`
`
`INTRODUCTION
`Recently, a new demand for demultiplexed formats
`has arisen in the seismic industry due to the
`utilization of minicomputers in digital field recording
`systems, and because of a growing need
`to
`standardize an acceptable data exchange format.
`In 1973 a subcommittee of the SEG committee on
`Technical Standards was organized
`to gather
`information and develop a nine-track, ½-inch tape,
`demultiplexed
`format
`for
`industry acceptance.
`Guidelines set for this new format were based on
`prior work and on the SEG Exchange Tape Format
`(Northwood et. al, 1967). As a result of the
`subcommittee’s effort based on suggestions from
`industry personnel,
`the following demultiplexed
`format recommendations are made.
`The present SEG Exchange Tape Format is often
`referred to as the SEG “Ex” Format. Because of this,
`it is recommended that the new demultiplexed format
`be designated the “SEG Y Format.” The Technical
`Standards committee has elected to withdraw support
`of the SEG “Ex” Format.
`The SEG Y Format was developed for application to
`computer field equipment and in the present data
`processing center with flexibility for expansion as
`new ideas are introduced. Current information for
`standardization is placed in the “fixed” portion of the
`format, while new ideas can be added to the
`unassigned portions later as expansion becomes
`necessary.
`
`
`1 1975 Society of Exploration Geophysicists. All rights reserved.
` This report is the work of the Subcommittee on Demultiplexed Field Tape Formats of the SEG
` Technical Standards Committee. Manuscript received by the Editor October 7, 1974.
`2 Subcommittee Chairman, Teledyne Exploration, Houston. Tex. 77036.
`3 Subcommittee Member, Gulf Research & Development Co., Houston, Tex. 77036.
`4 Subcommittee Member, Texas Instruments Inc., Houston, Tex. 77001.
`
`1 of 10
`
`WesternGeco Ex. 1012, pg. 1
`
`

`
`
`
`2.
`
`Fig. 1. Recommended demultiplexed format.
`
`Notes:
`1.
`Preamble-Proceeds each of the 45 blocks within the reel identification header and. each trace data block when 1600 bpi PE is
`used. Consists of 40 all-zero bytes followed by one all-ones byte.
`Postamble-Follows each of the 45 blocks within the reel identification header and each trace data block when 1600 bpi PE is
`used. Consists of one all-ones byte followed by 40 all-zero bytes.
`3.
`Interblock Gap (IBG)-Consists of 0.6" nominal, 0,5" minimum.
`4. End of file (EQF)-Consists of an IBG followed by:
`a) PE tape mark having 80 flux reversals at 3200 fci in bit numbers F,0,2,5,6, and 7. Bits 1,3, and 4 are dc-erased, or
`b) NRZI tape mark having two bytes with one bits in bit numbers 3,6, and 7 separated by seven all-zero bytes
`5. PE Identification Burst-Consists of 1600 flux reversals per inch in bit number P; all other tracks are erased.
`5) Definitions:
`d) PE identification burst – Consists of 1600
`a)
`Interblock gap (IBG) – Consists of erased
`flux reversals per inch in bit number P with
`tape for a distance of 0.6 inches nominal,
`all other traces DC erased. This burst is
`0.5 inches minimum.
`written beginning at least 1.7 inches before
`b) End of file (EOF) – Consists of the 800-bpi
`the trailing edge of the beginning of tape
`NRZI tape mark or the 1600-bpi tape mark
`(BOT) reflective marker and continuing past
`character, as appropriate, preceded by a
`the trailing edge of the marker, but ending at
`standard IBG.
`least 0.5 inches before the first block.
`c) Erased tape – The tape is magnetized, full
`e) Block – Continuous recorded information,
`width, in a direction such that the rim end of
`preceded and followed by a standard IBG.
`the tape is a north-seeking pole. The read-
`In PE (1600 bpi), a preamble precedes each
`back signal from such an area shall be less
`block and a postamble follows each block.
`f) Preamble – Consists of 41 bytes, 40 of
`than 4 percent of the average signal level at
`3200 flux reversals per inch.
`which contain zero bits in all tracks; these
`
`2 of 10
`
`WesternGeco Ex. 1012, pg. 2
`
`

`
`
`2-A EBCDIC CARD IMAGES Free form coding, left justified – 40 card images, 80 bytes per card, card image
`numbers 23-39 unassigned, for optional information.
`
`are followed by a single byte containing one
`bits in all tracks.
`g) Postamble – Consists of 41 bytes of which
`the first byte contains one bits in all tracks;
`it is followed by 40 bytes containing zero
`bits in all tracks.
`h) Two’s complement – Positive values are the
`true binary number. Negative values are
`obtained by inverting each bit of the positive
`binary number and adding one (1) to the
`least significant bit position.
`6) The seismic reel
`the reel
`into
`is divided
`identification header and the trace data blocks.
`The reel identification header section contains
`identification information pertaining to the entire
`reel and is subdivided into two blocks, the first
`
`containing 3200 bytes of EBCDIC card image
`information (equivalent of 40 cards) and the
`second consisting of 400 bytes of binary
`information. These two blocks of the reel
`identification header are separated from each
`other by an IBG. Each trace data block contains
`a trace identification header and the data values
`of the seismic channel or auxiliary channels.
`The reel identification header and the first trace
`data block are separated by an IBG.
`7) Each seismic-trace data block is ungapped and is
`written in demultiplexed format with each trace
`data block being separated from the next by an
`IBG. The last trace data block on the reel is
`followed by one (or more) EOF>
`
`3 of 10
`
`WesternGeco Ex. 1012, pg. 3
`
`

`
`8) When recorded 800 bpi (NRZI), the first block
`of the reel identification header begins at least
`3.0 inches past the trailing edge of the BOT
`marker.
`9) The following conventions pertain to the reel
`and trace identification headers:
`a) All binary entries are right justified. All
`EBCDIC entries are left justified.
`b) All times are in milliseconds with the
`exception of the sample interval which is
`designated in microseconds.
`c) All frequencies are in hertz.
`
`d) All frequency slopes are in dB/octave.
`e) All distances (lengths) are in feet or meters,
`and these systems are not mixed within a
`reel. The distance or measurement system
`used is specified in card image 7 and in
`bytes 3255-3256 of the reel identification
`header.
`f) A scaler may be applied to certain distance
`measurements where greater precision is
`required. See bytes 69-70 and 71-72 of the
`trace identification header.
`g) The energy source and geophone group
`
`Fig. 2A. Reel identification header. Part 1, the EBCDIC card image block.
`
`
`
`4 of 10
`
`WesternGeco Ex. 1012, pg. 4
`
`

`
`coordinates designated in bytes 73-88 of the
`trace identification header can be measured
`in either length or latitude and longitude.
`The measurement unit used is specified in
`bytes 89-90 of the trace header. For the
`latitude/longitude system,
`the coordinate
`values are expressed in seconds of arc.
`h) All velocities are in feet per second or
`meters per second, and these units are not
`mixed within a reel.
`i) Elevation is represented by “+” above “—“
`below mean sea level.
`10) The binary coded information convention is
`defined in Figure 1-C.
`
`
`DESCRIPTION OF REEL IDENTIFICATION HEADER
`The reel identification header (Figure 2) consists of
`3600 bytes and is divided into two parts:
`1) The card image EBCDIC block (3200) bytes—
`40 cards equivalent) followed by an IBG.
`2) The binary coded block (400 bytes) followed by
`an IBG.
`The EBCDIC part of the reel header describes
`the data from a line of shotpoints in a fixed specified
`format consisting of 40 card images with each image
`containing 80 bytes.
` All unused card
`image
`characters are EBCDIC Blank. Card image numbers
`23 through 39 are unassigned for optional use. Each
`card image should contain the character “C” in the
`first card column. Each 80 bytes would yield one
`line of format print to produce the form shown in
`Figure 2-A.
`The binary coded section of the reel header
`consists of 400 bytes of information common to the
`seismic data on the related reel as shown in Figure 2-
`B. There are 60 bytes assigned; 340 are unassigned
`for optional use.
`
`
`There are certain bytes of information that may
`not apply to a particular recording or processing
`procedure. It is strongly recommended that bytes
`designated with an asterisk (*) in Figures 2-B and 3-
`E always contain the required information
`The data in the reel identification header could
`be printed and edited prior to the actual input of
`seismic data for processing. A complete header
`listing of both the EBCDIC and binary parts would
`accompany an exchange tape and serve as a table of
`contents and summary of specifications for that reel*
`of seismic data. No more than one line of seismic
`data is permitted on any one reel. Additional reels
`would be used for long lines, and each reel must start
`with a reel identification header.
`
`
`DESCRIPTION OF THE TRACE DATA BLOCK
` Each trace data block (Figure 3) consists of a
`fixed 240-byte trace identification header and the
`seismic trace data block is separated from the next by
`an IBG. The trace header is written in binary code
`(refer to Figure 1-C for the binary code information)
`and is detailed in Figure 3-E.
` The trace data samples can be written in one of the
`four data sample formats described in Figures 3-A, 3-
`B, 3-C, and 3-D. The trace data format for each reel
`is
`identified
`in bytes 3225-3226 of
`the reel
`identification header. Only one data sample format is
`permitted within each reel.
` Figure 3-A details a 32-bit, floating point format in
`which each data value of a seismic channel is
`recorded in four successive bytes, in IBM compatible
`floating point notation as defined in IBM Form GA
`22-6821.
` The four bytes form a 32-bit word consisting of the
`sign bit QS, a seven-bit characteristic QC, and a 24-bit
`fraction QF. QS indicates signal polarity and is a one
`for a negative value. QC signifies a power of 16
`expressed in excess 64 binary notation allowing both
`negative and positive powers of 16 to be represented
`by a true number. QF is a six hexadecimal digit (24
`amplitude recovery can be described in the binary
`bit) number with a radix point to the left of the
`significant digit. The data value represented by a
`floating point number is
`
`
`QS X 16 (O C—64) X QF
`
`
`Figure 3-B details a 32-bit, fixed point format and
`each data value of a seismic channel is recorded in
`four successive bytes. This format consists of a sign
`bit QS (one represents negative) and 31 data bits QD
`with a radix point at the right of the least significant
`digit.
`
`Figure 3-C represents a 16-bit, fixed point format,
`and each data value of a seismic channel is recorded
`in two successive bytes. This format is similar to
`figure 3-B except there are 15 data bits QD.
`
`Figure 3-D represents a 32-bit, fixed point format
`with gain values. The first byte of this format is all
`zeros. The second byte provides eight available gain
`bits 20 through 27 . The last two bytes are identical to
`Figure 3-C.
`In all four data formats, the channel or trace data
`should represent the absolute input voltage at the
`recording instrument. The 32-bit, floating point field
`format defined as the SEG C (Meiners et al, 1972)
`comprehends the input voltage level. The fixed point
`formats 3-B and 3-C require a trace weighting factor
`
`5 of 10
`
`WesternGeco Ex. 1012, pg. 5
`
`

`
`Adoption and use of this format will save substantial
`sums of money in computer time and programming
`effort in the future.
`
`
`
`ACKNOWLEDGEMENTS
`
`
`Grateful appreciation goes to many companies and
`individuals for their suggestions at the start of the
`subcommittees' work
`and
`for
`their
`final
`recommendations. We are also for the assistance of
`Fred Tischler, Texas Instruments, who was the
`original subcommittee chairman.
`
`
`
`
`REFERENCES
`Meiners E. P., Lenz, L. L., Dalby, A. E., and
`Hornsby, J. M., 1972, Recommended standards for
`digital tape formats: Geophysics, v. 37, p. 36-44.
`Northwood E. J., Wisinger, R. C., and Bradley J. J.,
`1967, Recommended standards for digital
`tape
`formats: Geophysics, v. 32, p. 1073-1084.
`
`(trace identification header, bytes 169-170), defined
`as 2-n volts for the least significant bit, to comprehend
`the absolute input voltage level.
`
`In cases where processing parameters such as
`amplitude
`recovery are present,
`the
`type of
`amplitude
`recovery can be described
`in
`the
`appropriate reel identification header sections, and
`the algorithm described in the unassigned portions.
`
`
`
`CONCLUSION
`Individual oil companies and contractors may be
`convinced of their own format's merits, but the use of
`this recommended exchange demultiplexed format
`must be given serious consideration in order to
`achieve some level of industry standardization. Such
`thought and many suggestions from users have been
`utilized in establishing a flexible format that yields
`specifics and can be used by all companies in the in-
`dustry.
`
`
`
`6 of 10
`
`WesternGeco Ex. 1012, pg. 6
`
`

`
` 32 Bit Floating
`Point Format
`
`
`Sample Code=1
`3-A
`
`32 8it Fixed
`Point Format
`
`
`Sample Code=2
`3-B
`
`16 Bit Fixed
`Point Format
`
`
`Sample Code=3
`3-C
`
`32 Bit Fixed
`Point Format
`with Gain Values
`
`Sample Code=4
`3-D
`
`
`
`
`
`NOTE: Least significant bit is always in bit position 7 of byte 4 (or byte 2 for 3-C).
`
`
`QS = Sign bit
`QC = Characteristic
`QF = Fraction
`QD = Data bits
`
`FIG. 3A-D. Trace data block. Four data sample options.
`
`7 of 10
`
`WesternGeco Ex. 1012, pg. 7
`
`

`
`2-B. BINARY CODE-Right justified
`
`Byte Numbers
`
`3201-3204
`3205-3208
`3209-3212
`3213-3214
`
`
`
`Description
`
`*Strongly recommended that this information always be recorded.
`
`8 of 10
`
`Job identification number.
`Line number (only one line per reel).
`Reel number.
`Number of data traces per record (includes dummy and zero traces inserted to fill out the record or common
` depth point).
`Number of auxiliary traces per record (includes sweep, timing, gain, sync, and all other nondata traces).
`Sample interval in µsec (for this reel of data).
`Designated in microseconds to
`
`
`
`
`
`accommodate sample intervals less
`Sample interval in µsec (for original field recording).
`than one millisecond.
`Number of samples per data trace (for this reel of data).
`Number of samples per data trace (for original field recording).
`
`Data sample format code:
`1 = floating point (4 bytes)
`
`2 = fixed point (4 bytes.)
`Auxiliary traces use the same number of bytes per sample. (4 bytes)
`CDP fold (expected number of data traces per CDP ensemble).
`3 = single fold continuous profile
`
`Trace sorting code:
`
`1 = as recorded (no sorting)
`4 = horizontally stacked
`
`2 = CDP ensemble
`
`1 = no sum, 2 = two sum, ...., N = N sum (N = 32,767)
`
`3 = fixed point (2 bytes)
`4 = fixed point w/gain code
`
`
`1= linear
`2= parabolic
`
`
`
`
`
`
`
`3 = exponential
`4 = other
`
`
`Vertical sum code:
`Sweep frequency at start.
`Sweep frequency at end.
`Sweep length (msec).
`
`Sweep type code:
`
`
`
`Trace number of sweep channel.
`Sweep trace taper length in msec at start if tapered (the taper starts at zero time and is effective for this length).
`Sweep trace taper length in msec at end (the ending taper starts at sweep length minus the taper length at end).
`Taper type:
`
`1 = linear
`
`
`3 = other
`
`
`
`2 = cos2
`2 = yes
`
`
`Correlated data traces: 1 = no
`
`2 = no
`
`
`
`Binary gain recovered:
`1 = yes
`3 = AGC
`
`
`
`Amplitude recovery method:
`1 = none
`4 = other
`
`
`
`
`2 = spherical divergence
`2 = feet
`
`Measurement system:
`1 = meters
`
`Impulse signal
`
`1 = Increase in pressure or upward geophone case movement gives
`
`
`
` negative number on tape.
`Polarity
`
`
`2 = Increase in pressure or upward geophone case movement gives
`
`
`
` positive number on tape.
`Vibratory polarity code:
`Seismic signal lags pilot signal by:
`
`1
`=
`
`337.5° to 22.5°
`
`2
`=
`
` 22.5° to 67.5°
`
`3
`=
`
` 67.5° to 112.5°
`
`4
`=
`
`112.5° to 157.5°
`
`5
`=
`
`157.5° to 202.5°
`
`6
`=
`
`202.5° to 247.5°
`
`7
`=
`
`247.5° to 292.5°
`
`8
`=
`
`292.5° to 337.5°
`Unassigned – for optional information.
`
`
`
`
`*
`*
`*
`
`*
`*
`
`
`*
`
`*
`
`*
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`3215-3216
`3217-3218
`
`
`3219-3220
`3221-3222
`3223-3224
`3225-3226
`
`3227-3228
`3229-3230
`
`3231-3232
`3233-3234
`3235-3236
`3237-3238
`3239-3240
`
`
`3241-3242
`3243-3244
`3245-3246
`3247-3248
`
`
`3249-3250
`3251-3252
`3253-3254
`
`
`3255-3256
`3257-3258
`
`
`
`
`
`
`3259-3260
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`3261-3600
`
`
`
`
`
`WesternGeco Ex. 1012, pg. 8
`
`

`
`Byte
`Numbers
`1 - 4
`
`5 - 8
`9 –12
`13-16
`17-20
`
`21-24
`25-28
`29-30
`
`
`
`
`31-32
`
`33-34
`
`35-36
`37-40
`
`41-44
`
`45-48
`49-52
`53-56
`57-60
`61-64
`65- 68
`69-70
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`71-72
`
`
`
`
`73-76
`
`
`
`77-80
`
`
`
`81-84
`
`
`
`85-88
`
`89-90
`
`91-92
`
`93-94
`
`95-96
`
`97-98
`99-100
`101-102
`103-104
`
`
`number one.
`
`Description
`
`
`
`* Trace sequence number within line--numbers continue to increase if additional reels are required
`on same line.
`Trace sequence number within reel--each reel starts with trace
`* Original field record number.
`* Trace number within the original field record.
`Energy source point number--used when more than one record occurs at the same effective
`surface location.
`CDP ensemble number
`Trace number within the CDP ensemble--each ensemble starts with trace number one.
`* Trace identification code:
`7 = timing
`
`1 = seismic data
`4 = time break
`8 = water break
`
`2 = dead
`
`5 = uphole
`9---- N = optional use
`
`3 = dummy
`
`6 = sweep
`(N = 32,767)
`
`
`Number of vertically summed traces yielding this trace. (1 is one trace, 2 is two summed traces,
`etc.)
`Number of horizontally stacked traces yielding this trace. (1 is one trace, 2 is two stacked traces,
`etc.)
`Data use: 1 = production. 2 = test.
`Distance from source point to receiver group (negative if opposite to direction in which line is
`shot).
`Receiver group elevation; all elevations above sea level are positive and below sea level are
`negative.
`Surface elevation at source.
`Source depth below surface (a positive number).
`Datum elevation at receiver group.
`Datum elevation at source.
`Water depth at source.
`Water depth at group.
`Scaler to be applied to all elevations and depths specified in bytes 41-68 to give the real value.
`Scaler = 1, +10, +100, +1000, or +10,000. If positive, scaler is used as a multiplier; if negative,
`scaler is used as a divisor.
`Scaler to be applied to all coordinates specified in bytes 73-88 to give the real value. Scaler = 1,
`+10, +100, +1000, or +10,000.
`If positive, scaler is used as a multiplier: if negative, scaler is used as divisor.
`Source coordinate - X.
`
`If the coordinate units are in seconds of
`
`
`
`
`arc, the X values represent longitude and
`Source coordinate - Y.
`
`the Y values latitude. A positive value
`
`
`
`
`designates the number of seconds east of
`Group coordinate - X.
`
`Greenwich Meridian or north of the equator
`
`
`
`
`and a negative value designates the number
`Group coordinate - Y.
`
`of seconds south or west.
`Coordinate units: 1 = length (meters or feet). 2 = seconds of arc.
`Weathering velocity.
`Subweathering velocity.
`Uphole time at source.
`Uphole time at group.
`Source static correction.
`Group static correction.
`Total static applied. (Zero if no static has been applied,)
`
`FIG. 3E. Trace identification header written in binary code.
`
`9 of 10
`
`WesternGeco Ex. 1012, pg. 9
`
`

`
`
`Byte
`Numbers
`105-106
`
`107-108
`
`109-110
`
`111-112
`113-114
`115-116
`117-118
`119-120
`
`121-122
`123-124
`125-126
`127-128
`129-130
`131-132
`133-134
`135-136
`137-138
`139-140
`141-142
`143-144
`145-146
`147-148
`149-150
`151-152
`153-154
`155-156
`157-158
`159-160
`161-162
`163-164
`165-166
`167-168
`169-170
`171-172
`173-174
`175-176
`177-178
`179-180
`
`
`181-240
`
`Digital Tape Format
`
`Description
`
`
`
`
`Lag time A. Time in ms. between end of 240-byte trace identification header and time break.
`Positive if time break occurs after end of header, negative if time break occurs before end of
`header. Time break is defined as the initiation pulse which may be recorded on an auxiliary trace
`or as otherwise specified by the recording system.
`Lag Time B. Time in ms. between time break and the initiation time of the energy source. May be
`positive or negative.
`Delay according time. Time in ms. between initiation time of energy source and time when
`recording of data samples begins. (for deep water work if data recording does not start at zero
`time.)
`brute time--start.
`Mute time--end.
`* Number of samples in this trace.
`* Sample interval in µsec for this trace.
`Gain type of field instruments: 1 = fixed. 2 = binary. 3 = floating point.
` 4 --- N = optional use.
`Instrument gain constant.
`Instrument early or initial gain (dB).
`Correlated: 1 = no. 2 = yes .
`Sweep frequency at start.
`Sweep frequency at end.
`Sweep length in ms.
`Sweep type: 1 = linear. 2 = parabolic. 3 = exponential. 4 = other.
`Sweep trace taper length at start in ms.
`Sweep trace taper length at end in ms.
`Taper type: 1 = linear. 2 = cos2. 3 = other.
`Alias filter frequency, if used.
`Alias filter slope
`Notch filter frequency, if used.
`Notch filter slope.
`Low cut frequency, if used .
`High cut frequency, if used .
`Low cut slope
`High cut slope
`Year data recorded .
`Day of year.
`Hour of day (24 hour clock)
`Minute of hour.
`Second of minute.
`Time basis code: I = local. 2 = GMT. 3 = other.
`Trace weighting factor--defined as 2-N volts for the least significant bit. (N = 0, 1, .... 32, 767.)
`Geophone group number of roll switch position one.
`Geophone group number of trace number one within original field record .
`Geophone group number of last trace within original field record.
`Gap size (total number of groups dropped).
`Overtravel associated with taper at beginning or end of line:
`I = down (or behind). 2 = up (or ahead).
`Unassigned—for optional information.
`
` Strongly recommended that this information always be recorded.
`FIG. 3E. Trace identification header written in binary code (cont.)
`
`10 of 10
`
` *
`
`WesternGeco Ex. 1012, pg. 10